Purification, Characterization and Partial Amino Acid Sequencing of Hydroxycinnamoyl-CoA: Tyramine N-(hydroxycinnamoyl)transferase from Tobacco Cell-Suspension Cultures

Eur. J. Biocheni. 235, 762-768 (1996) 0 FERS 1YY6

Purification, characterization and partial amino acid sequencing of two new aspartic proteinases from fresh flowers of Cynara cardunculus L. Paula VERiSSIMO', Carlos FARO', Arthur J. G. MOIR', Yingzhang IdN', Jordan TANG' and Euclides PIRES' I Departamento de Rioquimica, Faculdadc de Ciincias e Tecnologia, Universidade de Coimbra, Portugal Department of-Molccular Biology and Hiotechnology, University of Shcffield, U K ' Protein Studies Program, Oklahoma Medical Research Foundation and Ilcpartment of Biochemistry and Molecular Biology, Univcrsity of Oklahoma Health Scicnce Center, USA (Received 31 October 1995) - EJB 05 179914

Two new aspartic proteinases have been isolated from stigmas of the cardoon Cyncrrn curdunculus L. by a two-step purificalion procedure including extraction at low pH, gel filtration on Superdex 200, and ion-exchange chromatography o n Mono Q. To follow the conventional nomenclature for aspartic proteinases, we have named these proteinases cardosin A and cardosin B. On SDS/PAGE, cardosin A migrated as two bands with apparent molecular masses of 31 000 Da and 15000 Da whercas the chains of cardosin B migrated as bands of 34000 Da and 14000 Da. The partial amino acid sequences of the two cardosins revealed that they are similar but not identical, and that they differ horn the previously reported cardoon proteinases named cynarases, which were assumed to be derived from a common precursor. Although thc cardosins show somc degree of similarity to each other, we could detect no immunological crossreactivity between theni. Both cardosins were active at low pH and were inhibited by pepslatin, with K , values of 3 nM for cardosin A and 1 nM for cardosin B, jndicaring that they belong to the class of aspartic proteinases. Significant differences between thc two enzymes were also found for thc K,,,,/k,,,values for the hydrolysis of two chromophoric synthetic peptides. The active-sitc ionization constants, pK,, and pKC2,for cardosin A are 2.5 -C 0.2 and 5.3 2 0.2, whci-eas for cardosin R they are 3.73 10.09 and 6.7 5 0.1. The results herein described on the structural and kinetic properties of the cardosins indicate that they are the products of distinct genes which havc probably arisen by gene duplication. A scheme for the proteolytic processing of the two enzymes is also proposed. Ke.ywordx: Cynaru curdunculus L. ; aspartyl proteinascs; milk-clotting enzymes; cardosins.

Aspartic proteinases are a group of enzymes that share many features in terms of sequence, three-di rnensional structure and catalytic mechanism [I -31. They are widely distributed in nature and have important roles i n biological systems such as precursor protein processing (retroviral proteases), protein degradation (pepsin, cathepsin D and fungal proteases) and blood-pressure regulation (rcnin) (for reviews, see [3-51). Only a small number of aspartic proteinases have been isolated and partially characterised from plants [6- 131. These proteinases, in common with most other aspartic proteinascs, havc an acid pH optimum, are inhibited by pepstatin and preferentially cleave pcptide bonds between hydrophobic residues. Little is known about their biological functions, but it has been suggested that plant aspartic proteinases rrre involved in the hydrolysis of storage and intracellular protcins 114- 171. Recently, three aspartic proteinases from barley, rice and cardoon have been cloned and their amino acid sequences deduced [18-201. A uniquc featurc sharcd by all these enzymes is an cxtra scgnienl C'orwspondmw I O C. J. C. Faro, Departamento Hioqui mica, Universidade Coimbra, Apartado 3126, P-3000 Coirnbra, Portugal ~hbr6'vi~lticms. &ln/HCI, guanidine hydrochloridc; Phe(NOZ),p - n trophenylalanine; Ahx, 2-uminohexanoic acid ; Tos-PheC~H2C1,L- I-ptosylamino-2-phcnylethylchloromethane. E,zz2vinr.r. Aspartic proteinases (EC 3,423); cathcpsin L) (RC 3.4.23.5); chymosin (EC 3.4.23.4): endoproteinase Glu-C (EC 3.4.21.19); pepsin (EC 3.4.23.1); renin (EC 3.4.23.13; lrypsin (EC

3.4.21.4).

of about 100 amino acids which bears no sequence similarity with aspartic proteinnses of inamtnalian or microbial origins. The tlowers of cardoon (genus Cynnm) are traditionally used in Portugal for cheese making and their proteinases are among the few eiizymes from vegctal sources that have been used for this purpose. We have previously reported the isolation of a proteiniise from commercially available dried cardoon flowers 121 1. This two-chain enzyme was shown 1221 to cleave k-casein at the same pcptide bond (Phe 105-Met106) as chymosin. A more recent study 1231 reported thc purification of three milk clotting enzytncs from these flowers. These enzymes were named cynarascs and were assumed to derive froin a common precursor by diffcrent processing. In the prcsent work, two additional aspartic proteinases were isolated frorn fresh stigmas of a standard variety of Cynur~1curdunrulms .1, grown from selected seeds. An investigation of the structural iind kinetic properties of these enzymes indicates that they arc the pruducts of different genes and that they differ from the previously reported cynarases. To follow the descriptive nomenclaturc for other aspartic proteinases, we have named thcse new proteiniises ciirdosin A and cardosin B.

MATERIALS AND METHODS Materials. Frcsh tlowers of C. curduricdus L. were collectcd from plants grown from weds wpplied by the Botanical

Gardens of the University of Coinibra. Pepstatin A was obtained froin the Peptide Institute, Inc. Diazoacetyl-DL-norleucincmethyl ester, trifluoromethanesulfonic acid and the peptide Leu-SerPhe(NO,)-Ahx-Ala-Lcu-OMe (Ahx, 2-aininohexanoic acid; Phe(NO,), p-nitrophenylalaninc) were purchased from Sigma, was USA. The peptide Lys-Pro-Ala-Glu-Phe-Phe(N0,)-Ala-Leu synthcsiscd at Krebs Institute, University of Sheffield, UK. Enzyme assay, The proteolytic activity was assayed using thc synthetic peptide Leu-Ser-Phe(N0,)-Ahx-Ala-Leu-OMe as substrate [24]. Enzyme preparations wcrc incubated at 37 "C with 0.8 mM substrate i n 50 mM sodium acetate, pH 4.7, 0.2 M NaCI, 4% (Me)?SO, and the rate of hydrolysis of Phe(N0,)-Ahx was nionitored at 310 nm i n a Perkin Elmer Lambda 2 UV/Vis spectrophotometer using the operating software. A molar absorption coefficient of 990 mM/cm at 310 nm was used in the calculations [24]. Protein determination. Protein concentration was determined by the method of Katzenellenbogen and Dobryszycka [2S] using bovine serum albumin as standard. Enzyme purification. Stigmas ( 1 g) from fresh tlowers of C. oardunculus L. were ground in a mortar and pestle under liquid nitrogen. The ground tissuc was then homogenised in 5 in1 0 . 1 M citric acid, pH 3.0, and centrifuged at 12000 g for 10 min. The supernatant (4 ml) was applied to a HiLoad Superdex 200 column equilibrated and eluted with 25 mM Tris/HCI, pH 7.6 (buffer A), at a flow rate of 1.O ml/niin. Each peak of absorbance was collected iis a fraction and assayed for activity. The active fraction was applied to a Mono 0 HR S/S column, also equilibrated in buffer A. The protein was eluted with a linear gradient of NaCl (0-0.5 M) in buffer A at a flow rate of 0.75 m h i n and the protein peaks were collected and assayed for activity. Polyacrylamide gel electrophoresis. SDS/PAGE was performed in a BioRad Mini Protean I1 electrophoresis apparatus according to the method of Laemmli [26] or in a Pharmacia PhastSystem using 20% homogeneous gels as described in the rnanufacture manual. Separation of the chains. Reverse-phase HPLC'. Cardosins were reduced and alkylated with 4-vinylpyridine, essentialy as describcd in 1271. The cardosins (20 pg) wcrc dissolved in 40 pI alkylalion buffer 16 M guanidine hydrochloridc (Gdn/HCl), 0.5 M 'his-CI, 2 mM EDTA, pH 7.51 and 1 pl 1.4 M dithiothreito1 and incubatcd for 1 h at room temperature. To this mixture, 1 p1 4-vinylpyridine was added and, after 5 min, the reaction was stopped by the addition of 10 pI 1.4 M dithiothreitol. The S-pyridylcthylated enzyme was then separated into polypeptide chains by reverse-phase HPLC using a Vydac C, column (4 mmX250 mm) equilibrated in 20% acetonitrile in 0.1 % (by vol.) CF,CO,H. Elution was carried out with a gradient of acctonitrile (20-80%) containing 0.1 o/u (by vol.) CF,CO,H in the clucnt. The tlow rate was 1.5 ml/tnin and thc eluent was monitored continuously at 215 ntn. Gel filtration in thr presence (,f6 M GddHCI. A sample of cardosin (0.5 ml) was dialysed against 0.1 M sodium phosphatc, pH 6.5, containing 6 M Gdn/HCI and applied to a Pharmacia k l W100 column packed with Sephadex G-l 00 which had becn equilibrated with the same buffer. The chains were eluted with the equilibration buffer at a low flow rate and the aborbance was monitored continuously at 280 nni. The eluent containing the isolated chains was dialysed exhaustively against 0.05 M NH,HCO, rind lyophilised. CNBr and enzymic cleavage of cardosins. CNBr digestion/ CNBr cleiivage was carried out in 70% CH,CO,H (0.5 ml) and approximately 100 pg CNBr. After 30 rnin at room temperature, the reaction mixture was lyophilised and redissolved in 8 M urea, 2% SDS, 200 mM Tris/bicine, pH 8.0, and 2 mM 2-mer-

captoethanol. The CNBr-cleaved peptides were then scparntcd by SDS/PAC;E followed by electroblotting onto poly(viny1idene ditluoride) membranes. Eqynzic deavnge. The isolated chains of each cardosin were incubatcd with ~-1-y-tosylamino-2-phenylcthylchloromethanetreated trypsin (4%, by mass) in 25 mM TridCl, pH 8.5, 0.3 M NaCl for 20 h at 37°C. The digests were thcn applied to an HPLC Vydac C,, column (4 mmX250 inin) which was cquilibrated with 0.1 % CF,CO,H, and the cardosin fragments eluted by a gradient of acetonitrile (0-80%) in 0.1 % CF,C02H at a flow rate of 0.5 tnl/min. The digestions of the isolated chains of cardosins A and I3 with endoproteinase Glu-C: (V, protcasc) wcrc carried out in 50 mM NH,HCO,, pH 8.1, at 30°C for 10 h using un enzyme/substrate ratio of 1 :40. Thc pcptides produced were isolated by reverse-phasc HPLC, iis described above for tryptic peptides. Endoprotcinase Gly-C digestion was performed i n the satne buffer (the enzymelsubstrate ratio was 1 :SO) for 3 11 at room temperature. The Gly-C peptides were separated by SDS/PAGE followcd by electroblotting onto poly(vinylidene ditluoride) membranes. Sequence analysis. N-tertniniil amino ucid sequences were determined by Edman degradation using an Applied Biosysknis 473-A Protein Sequencer equipped with a narrow bore HPLC for identification of the phenylthiohydantoin-atnino acids. Antibody production and Western-blot analysis. The isolated 3'1-kDa chain of cardosin A (0.5 mg) WBS ctnulsificd with Freund's complete adjuvant und injected subcutaneously into New Zealand rabbits. A second injection was tnade 2 weeks later using the same amount of isolated chain emulsified with Frcund's incomplete adjuvant, and antiserum was prepared from blood takcn I week after this last in.jection. For Wcstcrn-blot analysis, the cardosin chains were separiitcd by SDS/PAGE on a 1 2 % polyacrylarnide gel and transferred to a poly(viny1idene difluoride) rnernbrane by electroblotling in 1 0 niM 3-cyclohexylarnino- 1 -propanesulfonic acid, 10% tnethanol, pH 1 1.0, at 500 mA for 1 h. The membrane was incubated i n a blocking solution ( 2 . 5 % skirnrned milk in 0.1 M NaCl/O. I M sodium phosphate, pH 7.5/Tween) for 45 min at room tcmpcrature, then incubated overnight with an 1 : 500 dilution of the rabbit serum against thc 31-kDu chain. The tnembranc was washed three times with 80 mM Na,liPO,, 20 mM NaH,PO,, 100 mM NaCI, pH 7.5 (NaCIP,), 0.1 % Tween for 1 0 tnin and incubated with goat anti-rabbit IgG conjugated to horseradish peroxidase a1 a 1 :SO0 dilution for 1 11. After washing the tnetiibraiie three times with NaCI/P,/'Tween for 10 min, the peroxidase aclivity wiis dcvclopcd with 3,3'-diaminobeiizidinc (0.01% in NaCVPJlivcen) in the presence of 0.1 % H,02. Kinetics studies. For kinetics studies. the cxpcrimcntal conditions were those above described for the enzyme assay cxcept that the rate of hydrolysis for substrates was nionitored in ii HP 8452A diode array spectrophototneter using the HP89531 A UV/ VIS operating sortware. A molar absorption coefficient of 1480 mM/cm at 300 n m was used in the calculations when the synthetic peptidc Lys-Pro-Ala-C;lu-Phe-Phe(NO,)-Ala-Lcu was used as subslrate. Thc cnzytne concentration was determined by active-site titration of cardosins A and B with pepslatin A. For the kinetic studies at different pHs, the bulfers were SO m M CH,CICO,Nii (pH < 3.3, SO mM CH2C0,Nii (pH 3.5-5.5) and 50 mM Bis/l'ris/Cl (pH > 5.5). The hydrolysis of the peptide bonds Phc(N0,)-Ahx and Phe-Phc(N0,) was confirmed by revcrsc-phase HPLC. The kinctics parameters were culculated froin the Iheweaver-Burk plot using an approprialed software. The inhibition constant for pepslatin was calculated according lo Grccn and Work [28].

764

Verissimo et al. (Eur: 1. Hinchpm. 2.j.5)

Table 1. Summary of the purification procedure of cardosin A and cardosin B from fresh stigmas of C. cardunculus L.

Fig.1. SDSPAGE on a Phastgel homogeneous 20 of the enzyme preparation at each step of the purification of cardosins. The gel was stained with Coomassie brilliant blue. Lane 1, extract of cardoon stigmas; lane 2, first (inactive) peak from chromatography on a Mono Q column (Fig. 2 B ) ; lane 3, purified cardosin A (from the second peak, Fig. 2 8 ) ; lane 4, purified cardosin B (from the third peak, Fig. 28).

0.5

g

4

B

0.25

S

4

1.25

x3

E

l

0

Tmc (mm)

Fig. 2. Purification uf cardosins A and B. (A) Gel filtration on Supcrdex 200 of the acidic extract of C. cnrdunculus L. Samples (4ml) of an acidic cxtract of fresh stigmas of C. carduncmlus 1,. wcrc applied to a HiLoad Superdex 200 column equilibrated with 25 mM TrislCI, pH 7.6. The column was eluted with the same bufler at a flow rate of 1 mllmin. The protease activity is associated with the peak indicated by an arrow. (B) chromatography on Mono Q of the active fraction isolated from Superdex 200. The active fraction from Superdex 200 was applied to a HR 5/5 Mono Q column equilibrated with 25 m M TrislC1, pH 7.6. The column was eluted with a linear gradient of-0.5M NaCl in the equilibration buffer at a flow rate of 0.75 mllmin.

RESULTS Purification of cardosins. The milk curdling entyme preparation from the flowers of cardoon is traditionally obtained by water extraction or by grinding dried flowers with coarse salt into a fine powder. As water extraction yields an enzyme prcparation with a pH of about 5.5, this pH was initially used to extract cardosins from fresh flowers of cardoon. However, cxtraction at pH 3.0 was found to yield an extract with higher activity than that obtained at pH 7.6 or pH 5.5. In addition, extraction at low pH eliminated most protein contaminants, as judged by SDSPAGE analysis where only the bands from cardosins were observed (Fig. 1, lane 2). Extraction at low pH waq followed by gcl filtration on Superdex 200 (Fig. 2A). The proteolytic activity was recovered from this column as a singlc peak with an yield

step

Protein

Specific activity

Total activity

Yield

mg

Unitslmg

IJnits

B

Acidid cxtract Superdex 200 Mono Q cardosin A cardosin B

723 4.14

16.31 20.88

119.5

100 72.3

2.1 s 0.6

7.96 91.8

86.36 17.12 54.76

60.2

of about 72%. Peaks eluted after the active fraction were found to contain non-protein material, as no electrophoretic band was observed on SDS/PAGE. This material had a yellow color and tended to be retarded on the column. The partially puritied preparation obtained upon gel filtration was finally fractionated into three peaks by ion-exchange chromatography on Mono Q (Fig. 2B). The first component eluted at 17 min had no proteolytic activity towards the synthetic peptide uscd as substrate. The sccond and third peaks contained active enzymes which werc naincd cardosin A and cardosin B. SDSffAGE of these two proteinases rcvealed that each produced two bands either in the presencc or in the absence of 2-mercaptoethanol with apparent molecular masses of 31 kDa and IS kDa, respectively, for cardosin A and 34 kDa and 14 kDa, respectively, for cardosin B (Fig. 1 ) . The inactive component recovered from the Mono Q column also gave two bands on SDSPAGE corresponding to apparent molecular masses of 31 kDa and IS kDa (Table 1 ; Fig. 1). Comparisvn of cardosins. N-terminal amino acid sequencing. The relationship between cardosins was first investigated by determining thc N-terminal amino acid sequcnce of each chain after separation by SDSPAGE and electrotransference to poly(vinylidene ditluoride) membranes (Fig. 3). The sequences of the two chains from cardosin A are different, but homologous, to thosc obtained for the corresponding chains from cardosin B, demonstrating that the two cardosins are products of differcnt genes. In contrast, the N-tcrminal sequences of the 31 kDa and 15-kDa chains of cardosin A are identical to those obtained for the chains recovered from the material identified as peak 1 in Fig. 2 B , further confirming that this material is related to cardosin A. Intt,rnal ctmino acid sequencing. In order to obtain internal amino acid scquence data, the polypetide chains of both cardosins were isolated by reverse-phase HPLC or by gel filtration in the presence of 6 M Cdn/HCI. Reverse-phase chromatography on a C, column of the alkylated cardosin A yielded two peaks (Fig. 4 A ) which were confirmcd to be the 15-kDa and 31-kDa chains by SDS/PAGE.Using the same separation method, cardosin B was also fractionated into two peaks, which were identified by SDSPAGE as the 1CkDa and 34-kDa chains, respectively. The elution times of the cardosin R chains wcre different fmm those of the equivalent chains of cardosin A. Alteiiiaiively, the chains of each cardosin were isolated by gel filtration on Sephadex G I 0 0 in the presence of 6 M Gdn/HCI. In contrast, the chains could not bc separated by gel filtration i n the presence of 6 M urea (results not shown), suggesting that a strong intcraction cxists between the two chains of cach cardosin. The isolated chains wcre then digested with CNBr, trypsin and V 8 protcase and the peptide fragments were subsequcntly sepiirated by reverse-phase HPLC on a C,, column or by SDS/ PAGE followed by electrotransference to a poly(viny1idene ditluoride) membranc for automated Edman degradation. By this

765

Verissimo et al. ( E u r J . / ~ i r x : / i c w i .235) Cyp

1

cA(30m)

(38(34kn)

Cyp CR

51

Cyp

101

CA

RDSDGELIAL

KNyllDAaYFG

EILIG-TPPQ KFTVIFDTGS SNLOWPSSKC A

* * G S A W f f T*DR*TD*** **PTQ' GSGIV** T*DR*TD*ff **PTQ**fffN*A***DTG* *D***** YFSVACLFHS KYRSTDSTTY KKNGKSAAIQ YGTGSISGFF SQDSVKLGDL +*S**Sff +*SF*TK*

I

2

I1 I

LEVKEQDFIE ATKEPGITFL AAKFDGILGL GFQEISVGDA VPVWYTMLPI'Q ****a* **DfTDW** **** **fA*K+ ****-ff

CB

C y p 201 EMGDVLIGDK TTGFCASGCA AIADSGTSLL AGTTTIVTQI NQAIGAAGV

CA

Cyp

CR

10

tttft**t*

CE 251

MSQQCKSL----------349 PSPMGESAVD CSSLSSMPNI AFTVEEKTFN I ~ J * D )TSSE+LQ** *NT**R**kV ?*I* **K*G

CB

(MXD)

SAfSr** *NGI*$**NT A**I*+*K

C y p 379 LSPEQYVLKV GEGATAQCIS GFTAMDVAPP HGPLWILGDV MGQYHTVFD

Ch CB

*T***SY +T**

+SP** * *

Cfp 429 YGNLRVGFAE A A

CA CB

+*K** * * *

Fig. 3. Alignment of the partial amino acid sequences of cardusins A (CA) and I3 (CB) with that derived from the sequence of a partial cUNA of cynarase [ZO]. The partial sequence of each cardosin was determined upon sequencing of several peprides originated from digestions 01- the isolated subunits with CNBr, trypsin, endo-Gly and endoClu as described in Materials and Merhods. Identical residues are indicated by *. The activc-site triads DTG and DSG are show in bold. The N-glycosylated site of the 15-kDa chain is underlined.

method, thc partial amino acid sequcnces of the chains of both cardosins were determined (Fig. 3). revealing a clear similarity between the sequences of the cardosins A and B chains. A Surthcr siinilarity is also apparent between the cardosins and the prcviously reported partial cDNA sequence of a cynnrasc (Fig. 3). Inivnurioblotting unu1ysi.r. In order to further investigate the relationship between the isolated cardosins, the purified 3 1 -kDa chain of' cardosin A was uscd to raise a rabbit antiserum. Immunoblotting analysis showed that cardosin B did not crossrcact with this antiserum (Fig. 5), indicating that there i s no immunoidentity between the two cnzymes. As expected, the 31-kDa chain of the material from the first peak on Mono Q also reacts with the antiserum (Fig. S), suggesting, therefore, that this prcttein i s closely related to cardosin A. Enzymic properties and kinetic studies. Cardosins are activc at pH 2-7 with maximal activities around pH 5.0. Both enzymes are stable at tcmperaturcs up to 60 "C. In dilute solutions, the cardosins tend to be adsorbed onto solid surfaces, so they can bc trmsfcrrcd quantitatively only in the presence of a carrier prorein such as serum albumin or k-casein. Both enzymes wcrc iiihibitcd by the specific aspartic proteinase inhibitors, diazoacetyl-norleucine methyl ester and pepstatin. Inhibition constants determined in the presence of pepstatin are 3 nM for cardosin A and 1 nM for cardosin B (Fig. 6 ; Table 2). The pH dependence of the kinetic parameters K,,, k,,, and kCJKmfor cardosins was investigated using the peptide substrate Lys-Pro-Ala-

0

Fig. 4. Separation of the constituent polypeptide chains of cardosin A and cardosin B by reverse-phase-HPI,C.The reduced and alkylated cardosin A was applied to a Vydac C, column ccliiilibrated with 20% acetonitrile it1 0.1 % CF,CO,H. The chains wcrc eluted wirh a lincar gradient ol acetonitrilc (20- 80% in 40 min) :it a flow rate ol 1.5 ml/ miti. (A) Pcak 1, IS-kDa chain; peak 2, 31-kDa chain. Purified cardosin B was separated into its subunits by reverse-phase HPLC using the same condirions as described above. ( B ) Peak 1, 14-kDa chain; peak 2. 34kDa chain.

Fig. 5. lmmunoblotting analysis of the relationship between cardosin A and cardosin B. Thc isolated enzymes were suhjcctcd to SDS/PAC;E on a 12.5 % polyacrylamide gel arid transferred by electroblottirig onto poly(vinylidcnc difluoride) membranes. An nntiscrum raised against the isolated 31-kDa chain ol cardosin A was uscd >is ii probe. The antigenicanti body complex wiis developcd with a horseradish-peroxidase-coil-jugated goat anti-i-abbit serum using 3.3'-di;iminobcnzidline as chromogenic suhsrrate. Lane 1, cardosin B ; lane 2, cardosin A ; lane 3, inactive cardosin; lane 4,acid extract.

766

Vcrissimo et al. (Bur: J. Biochmt. 235)

Pepstatin Concentretion (nM)

Fig. 6. Inhibition of cardosin A (A) and cardosin B (B) by pepstatin A. The assays wcre carried out in 0 1 M sodium acetatc, pH 4.7, using thc as substrate. The inhibition constant w d b calculatcd according to the nicthod o l Green synthetic peptide Lys-Pro-Ala-Glu-Phe-Phc(N0,)-Ala-Leu and Work [28].

Table 2. Kinetic constants for the hydrolysis of two synthetic peptides by cardosin A and B. Enzyinc

Lyx-Pro-Ala-Cilu-Phe-Phe(NOL)-A1a-Leu

Leu-Ser-Phc(N0,)-Ahx- Ala-Leu

-

Cardosin A Cardosin B

KO,,

k,

.It

niM

s-

I

0.64 ? 0.02 0.081 f 0.01

13.7 2 1.6 86.2 ? 10.3

....

...

.

k,,,JK,,

K,,,

kc,,,

kJKn

mM ' s '

inM

s-i

rnM

21.3 ? 1.27 1065.7 ? 53.4

0.108 C 0.086 0.11 2 0.012

55.63 2 6.12 89.4 ? 3.5

515.09 ? 56.7 808.3 ? 32.3

Is-'

Glu-Phe-Phc-Phe(N0,)-Ah-Leu for cardosin A (Fig. 7, A-C) and the peptide Leu-Ser-Phe(N0,)-Ahx-Ah-Leu-OMe for cardosin B (Fig. 7, D-F). The pH-dependence curves are bell shapcd iind the values determined for thc apparent active-site ionimtion constants pK,, and pK,, of the free enzyme are rcspectively 2.5 -t 0.2 and 5.3 2 0.2 for cardosin A and 3.73 ? 0.09 and 6.7 -t 0. I for cardosin B. DISCUSSION

2.01 2.0

.

.

3.0 4.0

I . ,

.

5.0

PH

6.0

7.0

2.0

,

3.0 4.0

,

,

5.0

6.0

,

I

7.0

PH

Fig. 7. pH dependence of the kinetic parameters of cardosin A (a-c) and cardosin B (d-f). The kinetic parameters were determined 11s described in Materials and Methods. The synthetic peptide Leu-SerPhe(N0,)-Ahx-Ala-Leu-OMe was used for cardosin A and Lys-Pro-AlaGlu-Phe-Phe(NO,)-Ala-Leu was used for cardosin B. The buffers were SO mM CH,CICO,Na (pH .1:3.S), 50 mM CTI,CO,Na (pH 3.5-5.5) and SO mM Bis/Tris/CI(pH > 5.5).

The tlower of cardoon is a rich source of aspartic proteinases. Indeed, these enzymes account for inorc than 60% of the total protein i n mature stigmas and, to thc bcst of our knowledge, they seem to be thc first example of highly abundant aspartic proteinases in higher plants. Previous studies from this and other laboratories have described the isolation of aspartic proteinases from dried flowcrs of cardoon 121, 231. In the present work, we descrihe the isolation of two new aspartic proteinases from fresh stigmas of C. cardunculus L. using a simple proccdurc in which thc protcinases are first extracted from stigmas at low pH, lhen purified by gel filtration using Superdex 200 followed by ion-exchange chromatography using Mono Q, the entire purification being achieved in just 2 h. The simplicity in this isolation procedure compares favourably with that for the isolation of the cynarases [23]. In the present procedure, fresh flowers wcrc uscd a s starting inatcrial, which should bc a better source of the enzymes because it avoids the chemical modification occurred during the drying process. We have observed that thc specificity and kinetic properties of the enzymes isolated from dried flowers are different from those isolated from fresh

767

Verissimo ct al. (EULJ. Hiochmi. 23.5)

Cardosin A

1 mature cardosin A

Cardosin B

YJ 1(

ma-c&oiJlB

Fig. 8. Schematic representation of the proteolytic processing of cardosins. The two cardosins are likely ro be produccd ;ib single chains which undergo proteolytic processing at the N-terminus by removal of the proscquence. Thc twu precursors are further activated by removal of an internal sequence which is known as the plant-asparric-proteinase-spccific sequcnce. This sequence bears no similarity to the mammalian and rnicrohial aspartic proteinases. Pro, prosegment; N-Doni, N-terminal domain; Pss, planl-aspartic-pro~ei nases-specific sequence; C-l)om. C-terminal domain; GI, N-glycosylated site.

tlowers (unpublished results), suggesting that the catalytic properties of these enzymes arc modified during the drying process. The use of fresh flowers as the starting material has revealed, furthermore, that the elution profile obtained by chromatography on Mono Q i s depcndent on the type of cardoon, which has allowed us to perform a screening on the proteinase composition of different species of Cynara 1291. The partial amino acid sequence data (Fig. 3) indicate that the two cardosins are the products of different genes. Although they show sonie degree of similarity, we cannot detect any inimunological cross-reactivity hetween them. Furthermore, there arc significant differences between the enzymic and kinetic properties of the ciirdosins. The partial sequence data (Fig. 3) also reveals that the cardosins are distinct gene products from the cynarascs. In addition, the latter proteinases are assumed to bc derived from a single precursor by different proteolytic proccssing 1231, whereas cardosins are clearly encoded by two distinct genes. Evidence presented in this paper clearly shows that the cardosins helnng unequivocally to thc family of aspartic proteinases. The isolated cardosins arc inhibited by the general aspartic proteinase inhibitors, pepstatin and diazoacetyl-uL-norleucinc tncthyl ester, and are active at acid pH. Like the majority of the other aspartic proteinases, cardosins preferentially cleave peptide bonds between residues with hydrophobic side chains. Using oxidised insulin B chain as substrate, we have previously shown that cardosin B has a broader specificity than cardosin A [301. Clear differences were also found between the two ellzymes concerning the kinetic values for the hydrolysis of the synthetic peptide Leu-Ser-Phe(NOZ)-Ahx-Ala-Leu-oMc. The k,,,,lK,,,value determined for cardosin B is in the same range as that of pepsin (1640 mM-' s ') whereas the value for cardosin

A i s similar to that for chymosin (25.6 m M - ' s ' I * Although the precise specificities of the cardosins will reyuirc the study of more substrates, the results neverthcless suggcst that thc general specificity of cardosin A is similar to that of chymosin, whereas that of cardosin B closely rcseinbles the specificily of pepsin 1301. Cardosins have also different active-site ionization constants, although the pK values for the free enzymes fall within the range of the active-site ionization constants determined for other aspartic proteinases. However, it is noteworthy to mention that the values for both cardosins ure closcr to those of the human inirnunodeficiency virus type I protease [ 31 1 than of pcpsin and rhizopuspepsin 1321. This inuy be due to the fact that, like rctroviral proteinases, cardosins have optitnal activity under mild acid conditions. Among the known plant aspartic proteinases, cardosins appear to have a p t l optimum somewhat higher than [hat of thc other enzymes. In gcneral, the pH optimum of plant aspartic proteinases ranges over 2 - 2 5 for those enzymes from carnivorous plants [331 to pH 3.7-4.0 for the barlcy proieinase [ 01. The pH optimum of cardosins may reklect their subcellular localization. Like the barlcy prnteinasc 1341 and the most closely related aspartic protcinases from itnirrials 13-51and yeast [ 361, cardosins are likely to be vacuolar enzymes. However, they arc probably involved in different proteolytic evcnk, and while cardosin B tnay take part in general protcin digestion, cardosin A may have ii function in a rriore spccific regulated process. Nearly all of the aspartic proteinam a1-e synthcsised cis sitigle-chain zyinogens and activated upon removal of the propcptide [l]. Aspartic proteinases such as cathepsin I3 and the barley protcinase, however, undergo further proteolytic processing from single-chain to two-chain enzymes. It is possible that the activation of cardosins also takes place during the biosynthesis and

768

Verissimo et al. (6ur: .I. Hioclzmt. 235)

intracellular processing of the enzymes. The sequence data presented here suggest that processing of the single-chain cardosin to produce the two-chain enzyme is likely to occur through the cleavage of two peptide bonds, resulting in the removal of the region whose sequence is specific for plant aspartic proteinases (Fig. 8). This is suggested by sequence alignment of the cardosins to the barley proteinase and the cynarase, as well as by thc amino acid composition of each chain. The three-dimensional structures of a large number of aspartic prnteinases are known [ 2 ] , and rcccntly a three-dimensional iiiodel for the barley proteinase was proposed. According to this model, the specific insert is present at the C-domain near thc surface between residues G239 and G243 1371. This sequence is located in a position that would probably generate a second loop over the active site, resulting in a hindrance to the access of substrate, and thus should be removed in order to render a fully active enzyme. The existence of an inactive form containing the plant-specific insert may therefore be an interesting stratcgy to overcome unwanted protcolysis in the cell and the search for such form in cardosins is currently under investigation in our laboratory.

16. Elpidina, E. N., Dunaevsky, Y. E. & Belozersky, M. A. (1990) Protein bodies from buckwheat seed cotyledons: isolation and char-

acterization, J. Exp. Hot. 41, 969-977. 17. Rodrigo, I., Vcra, P. & Van Loon, L. C. (1991) Degradation of tobacco pathogenesis-related proteins, Plant Physiol. 9.5, 61 6-622. IS. Runeberg-Roos, P., Tormakangas, K. &L Ostman, A. (1991) Prirnary structure of ti barley-grain aspartic proteinase. A plant aspartic proteinase rescmhling mammalian cathepsin D, Eur J. Bimt.llrm 202, 1021 -1027. 19. Asakura, T., WLttanabe, H., Gbe, K. & Arai, S. (1995) Kice aspartic proleinasc, oryzasin, expressed during seed ripening and germination, has a gene organization distinct from those of anitnal and microbial aspartic proteinases, Eur: .I. Riochem. 232, 77-83. 20. Cordeiro, M. C., Xuc, Z.-T., Pietrzak, M., Pais, M. S. & Brodelius, P.E. (1994) Isolation and characterization of a cDNA from flowers of C:vruuru cnr~iunculuscncoding cyprosin (an aspartic proteinase) and its use to study the organ-specilic expression of cyprosin, Plunr Mol. B i d 24, 733-741. 21. Faro, C., Alfacc, J. S. & Pires, E. V. (1987) Purification of ia protease from the flowers of C-vnurucurdurzculus Id., C i h . Hiol. 12, SG 201. 22. Fxo, C. J., Moir, A. 3. C. CQPires, E. V. (1992) Specificity o f a milk clotting enzyme extracted from the lhistle C:vnura C ~ d U n C U / L 6 . L. t : action on oxidiscd insulin and K-casein, Biotech. Lett. 14, 841 This work was mainly supported by JNICT, Portugal (contract 846. number PMCT/C/B10/909/90). Paula Verissirno is the recipient of a Phll 23. fIeimgartner, U., Pietrzak, M., Geertsen, R., Brodelius, P. PiU de lnvrstiga@o Cient@cu c Trcnoldgir:~ fellowship from J U ~ INwionul gueiredo, A. c‘. 62 Pais, M. s. (1990) Purification and partial char(JNICT), Portugal. acterization of milk clotting proteinases from flowers of Cynnru c~arrfunculusL., fhytochemistly 2.9, 1405-1 410. REFERENCES 24. Martin, P. (1984) Hydrolysis of the synthetic chrotnophoric hexapeptiile Leu-Ser-Phe(NO,)-Nle-Ala-l.eii-oMe catalyzed by 1. Tang, J. & Wong, R. N. S. (1987) Evolution in the structure and bovine pepsin A, Riochem. Hiophys. Acru 791, 28-36. function of aspartic proteinases. J. C’rll Rioche 25. Katzenellenbogcn, W. M. & Dobryszych, W. M. (1959) New 2. navies, D. R. (1990) The structure and function of thc aspartic promethod for quantitativc determination of serum proteins separated teinases, Annu. Rev. Biopltys. Chern. 19, 189-215. by p:ipcr electrophoresis, Clin. Chim. Acta 4 , 5157-522. 3, Dunn, B. M. (1992) Structure undfirnrtion qf tbe aspartic. protein26. Ltemmli, U . K. (1970) Cleavage of structural proteins during thc a . m (Dunn, B., ed.) Plenum Press, New York. assembly of the head of bacteriophage T4, Nature 227, 680-685. 4. Tang, J. (1977) Acid pmteases. structure, ,func!ion arid biolcigy 1.1.62 Yuan, P. (1Y87) Microscale alkylation with 4-vinylpyr27. Hawkc, (Tang, J., ed.) Plenum Press, New York. idine, Apl~liedHiosystem U S ~ T ’ S bullerin issue 28. 5. Kostka, V. (1985) Aspartic proteina.res trnd their inhihilors (Kostka, 28. Green, N. M. 62 Work, t.:. (1953) Pancreatic trypsin inhibitor: 2. V., ed.) Walter de Gruyter, Berlin. Reaction with trypsin, Biocbem. J. 1.54, 347-352. 6. Sarkkinen, P., Kalkkinen, N., Tilgmann, C., Siuro, J., Kervinen, J. & Mikkola, L. (1997,) Aspartic proteinase from barley grains is re- 29. Estevcc, C‘. L. (199s) Bstudo comparativo das carracteristicas bioquirriicas dos coalhos de Cywum r~arr1unculu.rL., Cynaru . r c d y lated to mammalian 1ysosomal cathcpsin D, PIanta (Berlin) 186, rriu.~L. e Cynaru humilis L., MSc Thesis, University of Coirnhra, 317-323. Portugal. 7. Belozersky, M. A., Sarbakanova, S. T. 62 Dunaevsky, Y. E. (1989) Aspartic proteinase from wheat sccds: isolation, properties and 30. Verissinio, P., Esteves, C., Faro, C. & Pires, E. (1995) The vegetable rennet of Cyiartr rzrdtmculus L. contains two proteinases with action on gliadin, Plunta (Berlin) 177, 321 -326. chymosin and pepsin-like specificities, Biotech. Left. 17, 621 -625. 8. Jloi, E., Shibata, n., Matoba, T. & YoneLawia, D. (1980) Characterization of pepstatin-sensitive acid protcase in resting rice seeds. 31. Ido, E., Ih n , H. I?, Kezdy, E J. & Tang, J. (1991) Kinetic studies of human immunodeficiency virus type 1 proteasc and ils activcAgric. B ~ c JChem. ~ . 44, 741 -747. site hydrogen bond mutant A28S, J. B i d . Chmnz. 266, 243599. Polanowski, A., Wilusz, 1:.Kolaczkowska, M. K., Wieczorck, M. & 24 366. Wilimowska-Pelc, A. (1985) Purification and charactcrization of 32. Lin, Y., k‘usek, M., Lin, X., Hartsuck, J. A,, Kezdy, E J. & Tang, J. aspartic protcinases from Cucumis sativus and Curcuhita maxima ( 1992) pH dependcnce ol’ kinetic parameters of pcpsin, rhizopusseeds, in Aspartic prutsinases and rhrir inhihitmi (Kostka, V., pepsin and thcir active-site hydrogen bond mutants, J. Biol. ed.) pp. 49-52, Walter de Gruyter, New York. C/wllem. 267, 18413- 18418. 10. Garg, G. K. & Virupaksha, T. K. (1970) Acid Proteasc from germinated sorghum : 1-purification and characterization of the enzyme, 33. Tokts. Z. A., Woon, W. C. & Chambers, S. M. (1974) Digestive enzymes secreted by the carnivorous plant Nepenihes macferlunri EUV:J. Binchem. 17, 4-12. L., Pltrniu (Berlin) 119, 39-46. 11. Kodrigo, I.. Vera, P. & Conejero, V. (1989) Dcgradration of tomato pathogcnesis-related proteins by an endogenous 37-kDa aspartyl 34. Runeberg-Roos, P., Kervinen, J., Kovaleva, V., Kaikhel, N. V. & Gal, S. (1994) The :ispartic proteinase of harley is a vacuolar endoprotcinase, Eur: J. Bincbcnt, 184, 663 -669. enzyme that processes probarley lectin in vitro, Plant P b p i o l . 12. Bourgeois, J. & Malek, L. (1991) Purification and characterization 10.5, 321 329. of the aspartyl proteinase from dry jack pine seeds, Seed Sci. RPS. 35. Barret, A. J , ( I 977) ProtPinases in rnurnninlirxn cdls und tissues, I , 139-147. pp. 209-248, Elsevier-North Holland, New York. 13. Salmia, M. A. (,1981)Proteinascs activities in resting and germinat36. Aminerer, Ci., Hunter, C. P., Rothnian, J. H.. Saari, G. C., Valls, L. ing seeds of Scots pine, Pinus sylvestris, Wtv.rio1. Pbnt. 53, 39-47. A. & Stevcns, T. H. (19x6) PEP4 gene of Scrcch~7rom,v~~vrrs (wcvi. 14. Juniper, B. E., Robins, K.J . & Joel, D. M. (1989) Tbe Carriivorous siac cncodes proteinase A, a vacuolar enzynic required for prophZt.9 (Juniper, B. E., Robins. R. J. & Joel, D. M., eds) pp. 189cessing of vacuolar prccursors, M d . Cell Hiol. 6 , 2490-2499. 207, Academic Press Inc., London. 15. D’Hondt, K., Bosch, D., Van Damme, J., Goethals, M., Vandekerck- 37. Citiruprasad, K., ‘Iormakangas, K., Kervinen, J. & Blundell, T. (1904) Comparative modelling of barley-grain aspartic proteinhove, J. & Krebbers, E. (1993) A n aspartic proteinasc present in iise: a: structural rationale for ohscrved hydrolytic specificiry, seeds cleaves Arabidopsis 2s albumin precursors in vitro, J. Ria/. FETkS I , P / ~3.52, . 131- 136. Cbem. 268, 20884-20891.